110W Flyback

[chimi]

Hi, this is related to the prior post I did today... in certain way I answered myself one of the questions and decided to go thru with the 110W flyback, but I think I'll need a little help qith the main input filtering and surge supression.
Here I attach the only off-line SMPS I commertialized for a while as it is, just as a power supply, for a casino machine that needed 19Vdc - 3.42A, so this is a 70W off-line flyback converter, that works pretty nice, but didn't need to meet any standards that I know of, but it was in a metalic cabinet, so there was no problem...
Now I need something more adequate for a machine that maybe is going overseas, so, maybe hasn't to meet all IEEE standards, but at least I want it to throw the least possible noise at low cost to the mains power.
So, I'd be really grateful if you take a look to the schematic I post and tell what would you change, mostly in the input filtering, which is almost non existent
Thanx!!

[Tahmid]

Hi,
You could insert 0.1uF 400v caps before and after the line filter, and across the CRAIL 100u reservoir cap. Insert an RCD snubber and/or transient/surge suppressor (transzorb/transil + UF diode[take a look at P6KE200]). You should use an auxilary power supply (an extra winding in the transformer) to fully power the UC3844.
Thanks.

[Tahmid]

One more thing,
You should decouple the VCC and VREF pins with 0.1uF ceramic caps placed as close as possible to the IC. This is very important for the long lasting of the controller.
Thanks.

[Dibley]

Tahmid,

Good points. I would leave the X2 cap in front of the bridge out though, I havent seemed to gain much in the 150K to 30MHz region from that one in shown filter.

Dibley.

[mflorin]

For a commercial product, an input filter is mandatory to meet standards specifications.
Knowing the switched input current waveform, the amplitude of the fundamental can be calculated. Then, depending on which standard you want to consider, the necessary input filter attenuation can be found.
http://www.alsdesign.fr/alsftp1/note...69-basso01.pdf

[star882]

Where's the input fuse and inrush protection? I'm sure they are required for UL compliance.

Maybe try spread spectruming the clock? A few resistors to modulate the clock frequency from the incoming AC waveform might be the easiest way to do it.

[chimi]

Hi there and thank you all!
Well, I know this does not meet any standards, as I told you before, but works nicely just under 72W, the question is: is this circuit suitable for 130W?

I avoided the use of RCD snubbers and auxiliary windings, instead I used a passive "non-dissipative" snubber that uses the filtered spikes from switching-off to feed the PWM IC, reducing the leakage inductance in the transformer and using the energy that would be dissipated across the RCD snubber's resistor as a power source for the logic, hoping this would not affect the logic with noise...

Regards!